Process metrics-oriented study on alkyl esters enzymatic syntheses using immobilized lipases for assertive process development: n -octyl acetate synthesis as a case-study

1 Introduction

Esters are substances of great industrial relevance to the cosmetic, food, pharmaceutical, and fuel sectors. Industrially, they are produced with catalysts of low selectivity and specificity, leading to the formation of byproducts and rigorous post-reaction purification. Enzyme-driven catalysis (biocatalysis) can overcome this bottleneck under mild reaction conditions, allowing for green and efficient syntheses (Teetz et al., 2024). However, the biocatalytic route for esters production should be translated into process-relevant conditions, considering the need for high yields and productivities starting from high substrate concentration media (Erdem and Woodley, 2024; Robescu and Bavaro, 2026; Domínguez de Maria, 2024). The high cost of biocatalysts – lipases, in immobilized form – hinders their large-scale adoption for low-moderate selling price products like most esters, even though the literature reports many successful applications of immobilized lipases in ester syntheses (Sousa et al., 2023).

Carboxylic acids and alcohols, in general, are mutually soluble, enabling ester syntheses in a solvent-free system (SFS). Solvent-free systems demonstrate several advantages from a process perspective, allowing high substrate/product concentrations in the media and consequently high reactor volumetric productivity (Sandoval et al., 2002). Unit operations to separate and recycle solvents are not required in SFS, representing process simplification and capital investment/operational costs saving. The literature describes that most immobilized lipases are stable in organic media with high substrate concentration in most cases (Klibanov 2001; Sandoval et al., 2002; Sousa R. R. et al., 2021). Mid- and long-chain acids and alcohols are hydrophobic and demonstrate discrete or no inhibition/inactivation potential on lipase activity, but, contrarily, short-chain acids and alcohols cause deleterious effects on lipases, as identified in many studies (Bélafi-Bakó et al., 2003; Nordblad and Adlercreutz, 2008; Foresti et al., 2008). These effects include possible distortions on the enzyme hydration layer leading to a reduction in conformational flexibility required to catalytic activity, drastic pH reduction on the hydration layer, and the inhibitory (competitive inhibition) effect itself. Thus, short-chain ester syntheses (from the reaction of a short-chain acid and/or short-chain alcohol) are frequently performed using hydrophobic solvents such as n-hexane/n-heptane or other hydrocarbons. These solvents reduce the concentration of inhibitors (short-chain acids and alcohols) and increase overall hydrophobicity, allowing proper lipase activity. In summary, mid/long-chain ester syntheses can easily be carried out in solvent-free systems achieving high product concentrations/productivity, whereas short-chain ester syntheses catalyzed by lipases generally require solvents achieving less favorable process metrics.

The lack of studies focused on process metrics calculations, modeling, and process simulation, and techno-economic assessments in this area makes it difficult an accurate identification of the technical and economic bottlenecks of biocatalysis applied to esters production, as well as their technical and environmental advantages on a quantitative basis, reducing the attractiveness of this green technology (Sousa et al., 2023). Additionally, industrial research is usually less disseminated than academic research, causing a gap of information that impairs a convergence towards biocatalysis development for commodities/specialties production (Domínguez de Maria, 2024). Most studies on enzymatic esterification dedicate efforts to reaction conversion and its optimization, but few of them include the calculation and optimization of process productivity, such as space-time yield (STY, or volumetric productivity), and catalytic productivity (Hauer, 2020; Remonatto et al., 2022a). For instance, high conversion of reagents into esters can be achieved using a stoichiometric excess of alcohol, but depending on the excedent quantity of alcohol, the downstream operations will be complicated, demanding several recovery/recycling steps, implicating additional costs. Another usually neglected aspect on reaction parameters definition is the biocatalyst loading, which has a direct correlation with the reaction time. As the biocatalyst cost (around 100–1000 USD kg^-1) is much higher than the most industrial-relevant ester cost (around 2–10 USD kg^-1), the biocatalyst loading should be as low as possible to achieve high catalytic productivity (mass of product obtained per mass of biocatalyst) in the process. These issues have been pointed out in general terms in some papers (Tufvesson et al., 2010; Tufvesson et al., 2011; Lima-Ramos et al., 2014; Gomes and Woodley, 2019; Woodley, 2019; Woodley, 2022; Woodley, 2025), but in enzymatic esterification studies, the process metrics calculation has little interest by academic groups up to now (Hauer, 2020; Remonatto et al., 2022a).

This paper aims to study ester syntheses catalyzed by immobilized lipases under a process metrics approach. We intend to obtain high conversions conciliated with high volumetric and catalytic productivity, which are fundamental metrics for techno/economic feasibility. The solvent-free syntheses of ethyl and octyl esters (acetate, caprate, and oleate) using lipase B from Candida antarctica immobilized (Novozym 435) were studied to establish the optimization paths that lead to favorable process conditions. The output data for the enzymatic n-octyl acetate synthesis in the lab-scale were used to process design, modeling/simulating study, and a simplified techno-economic assessment to identify the process drawbacks.

2 Materials and methods

2.1 Reagents and immobilized lipases

Capric acid, oleic acid, and n-octanol were purchased from Merck (Germany); acetic acid and ethanol were purchased from LabSynth (Brazil). All reagents were of analytical grade. Novozym 435 was kindly donated by Novozymes Brazil (currently Novonesis). Potassium hydroxide was used for acids titration, and isobutanol (both from LabSynth, Brazil) was used for the dilution of samples in gas chromatography analyzes.

2.2 Overall reaction procedure

Esterification reactions were performed by direct mixture of carboxylic acid (acetic acid or caprylic acid or oleic acid) and alcohol (ethanol or n-octanol) without any solvent. The syntheses were carried out during 4 h using acid:alcohol molar ratios 1:1 or 1:2, at 50 °C, and biocatalyst loading of 100 U (considering Novozym 435 esterification activity as 2904 U g^-1 in standard synthesis of ethyl oleate at 40 °C). The reactions were carried out in capped Erlenmeyer flask (25 mL), and a typical reaction starts using 0.01 mol of acid and an equivalent quantity of alcohol. Adjustments in molar quantity (a little more or less) were made depending on the reagents involved to maintain the total reagent mass close to 2.5–2.6 g. The biocatalyst loading tested (100 U) was equivalent to 1.3%–1.4% wt/wt total mass reagent in this condition, depending on the reagents involved. Each synthesis was studied in four different conditions: molar ratio acid:alcohol 1:1 without and with water removal (using molecular sieves directly added in the reaction media), and molar ratio acid:alcohol 1:2 without and with water removal. When molecular sieve was present in the reaction media (30% wt/wt total reagent mass), the reaction conditions were labelled as “wr” (water removal).

2.3 Monitoring esterification reactions performance

Esterification reactions were monitored by the reduction of the acidity index. The acid conversion was measured by titration of samples (60 µL aliquot dissolved in 5 mL ethanol) collected at the beginning of the reaction and at the end of a period. Titration was made using aqueous KOH solution (0.008 M). The ester concentration was calculated from the stoichiometry of the reaction and confirmed by gas chromatography (Agilent 7890D) using a SupelcoWax10 column (30 m × 0.53 mm x 0.25 µm). For short-chain reagents/esters analyzes, oven program was: 45 °C for 2 min, heated to 60 °C by 5 °C min^-1 and held for 2 min; heated to 150 °C by 5 °C min^-1 and held for 4 min; heated to 250 °C by 20 °C min^-1 and held for 1 min (total analysis time 29 min, 1 µL injection volume, split mode 2:1). For mid/long-chain reagents/esters analyzes, oven program was: 45 °C for 1 min, heated to 120 °C by 25 °C min^-1 and held for 2 min; heated to 250 °C by 50 °C min^-1 and held for 20 min (total analysis time 28.6 min, 0.3 µL injection volume, split mode 2:1).

2.4 Process metrics study

The process metrics studied include conversion of the reaction, product concentration, space-time yield (STY, volumetric productivity), specific yield (catalytic productivity), added-value index (AVI), and E-factor. The conversion result and ester concentration (mol or g L^-1) were used to calculate the catalytic productivity–mol (or g) product g biocatalyst^-1 – for each synthesis, as shown in Equations 1–6. Catalytic productivity per cycle was calculated using the data obtained in the first cycle, and total catalytic productivity was calculated using the quantity of product obtained in the first cycle multiplied by the number of cycles using the same biocatalyst loading.

C o n v e r s i o n = 1 − F i n a l A c i d i t y I n i t i a l A c i d i t y × 100 (1)

P r o d . C o n c e n t r a t i o n m o l L − 1 = m o l a c i d c o n v e r s i o n 100 R e a c t i o n v o l u m e (2)

S T Y m o l L − 1 h − 1 = P r o d u c t c o n c e n t r a t i o n R e a c t i o n t i m e h (3)

C a t a l y t i c P r o d u c t i v i t y m o l g − 1 = P r o d c o n c e n t r a t i o n × r e a c t i o n v o l u m e B i o c a t a l y s t m a s s g (4)

A V I = m a s s p r o d u c t × p r o d u c t s e l l i n g p r i c e ∑ m a s s r e a g e n t s × r e a g e n t c o s t (5)

E − f a c t o r = m w a s t e k g m p r o d u c t k g (6)

2.5 Process design and preliminary techno-economic assessment

The data obtained in lab-scale n-octyl acetate enzymatic synthesis was used to simulate an industrial-scale process in SuperPro Designer software (v13. Intelligen Inc., USA). The process design includes steps of storage and pumping of reagents (acetic acid and n-octanol), the reaction itself in a bubble-column reactor stirred by an ascending compressed air stream at atmospheric pressure, and purification steps using vacuum distillation columns, as demonstrated in Figure 1. The distillation columns were modeled in ChemSep software (v8.4, Harry Kooijman and Ross Taylor, USA), using the Gamma-Phi thermodynamic equilibrium models, the Redlich-Kwong equations of state (1949), and the UNIFAC thermodynamic activity coefficient model (1975). For simplification purposes, we considered that the reaction conversion and reuse capacity of Novozym 435 remains the same when scaled up from lab-scale to industrial-scale simulation. The authors recognize that this is an optimistic approach; high-accuracy data may certainly be obtained in scaling-up studies, which are out of the scope of this work. Based on this limitation, the output data obtained here may be considered as a starting point for techno-economic drawbacks identification. The premises adopted for the mass and energy balances calculation, equipment sizing estimatives, CAPEX and OPEX estimates, and techno-economic assessment (Net Present Value, Internal Rate of Return, and Payback time) are shown in Table 1.

TABLE 1. Premises and reference values for process design and preliminary techno-economic assessment of octyl acetate continuous production (7920 h yr^-1) using Novozym 435 as biocatalyst.

3 Results and discussion

In the initial stages of process development, atom economy and carbon mass efficiency are important parameters for defining the production route to be chosen (Lima-Ramos et al., 2014). In general, esterification or transesterification is applied to the esters production. However, the esterification reaction has high atom economy and high carbon efficiency when compared to the equivalent transesterification reaction, that generates glycerol as by-product (3 carbons are lost from the reagents) while esterification generates only water as by-product. Due to this and to the high specific catalytic activity of lipases, the E-factor metric – the amount of waste created per kg of product produced – is close to zero if quantitative conversion in equimolar condition is achieved (>99%) (Lima-Ramos et al., 2014; Sheldon, 2017). Thus, for a given ester production the esterification route shows more favorable metrics when compared to transesterification.

In laboratory-scale development, reaction conversion, the final concentration of product, space-time yield (STY), and (bio)catalytic productivity are key parameters for achieving optimized reaction conditions (Konzock and Nielsen, 2024; Lima-Ramos et al., 2014; Woodley, 2022). At the industrial level, where high capital investments are involved, the economic metrics which are directly correlated with the process’s ones are predominant and guide the targets for other operational indicators during process development (Towler and Sinnot, 2013; Turton et al., 2018; Meissner and Woodley, 2022). Thus, the following results obtained in enzymatic ester syntheses considered both aspects – operational and economic metrics.

The results shown in Table 2 indicate that mid/long-chain ester syntheses achieved higher conversions than the short-chain ones. The same trend is observed when compared one synthesis that involved one short-chain reagent with one synthesis with only mid/long-chain reagents, for instance, ethyl oleate and octyl oleate (78.5% and 87.3% conversion in 1:1 wr, respectively). The synthesis of esters derived from mid/long-chain acids and alcohols generates a hydrophobic reaction medium, in which the hydration layer of lipases (responsible for the flexibility necessary for catalytic activity) is not significantly affected. Furthermore, mid/long-chain acids and alcohols do not exhibit an intense inhibition effect on lipases, as evidenced by kinetic modeling studies (Alves et al., 2016; Chowdhury and Mitra, 2015; Foresti et al., 2008; Lima et al., 2018; Serrano-Arnaldos et al., 2020; Silva et al., 2021). Additionally, the hydrophobicity of the medium increases during the ester formation, causing a partitioning of the formed water, which limits the occurrence of the hydrolysis (Sandoval et al., 2002; Sousa R. R. et al., 2021). Thus, the syntheses of mid/long-chain esters in equimolar conditions without water removal, i.e., without equilibrium shift strategies, achieve conversions close to or higher than 80%, as shown in Table 2 for the syntheses of ethyl caprate, octyl caprate, ethyl oleate, and octyl oleate. The same is attested in many studies with different immobilized lipases (Garcia et al., 1999; Gawas et al., 2016; Rocha et al., 1999; Sandoval et al., 2002; Serrano-Arnaldos et al., 2020; Sousa et al., 2020; Sousa R. R. et al., 2021; Yadav and Thorat, 2012). The reaction media of an esterification involving only mid/long-chain reagents have favorable thermodynamics and favorable conditions for lipase’s activity.

TABLE 2. Conversion results obtained in four different reaction conditions in the enzymatic syntheses of alkyl esters–ethyl acetate, octyl acetate, ethyl caprate, octyl caprate, ethyl oleate, and octyl oleate. All syntheses were catalyzed by Novozym 435 (100 U biocatalyst loading) and carried out in stoppered flasks (25 mL) under orbital stirring (200 rpm), at 50 °C, 4 h reaction time.

Contrarily, the syntheses involving short-chain acids and short-chain alcohols demonstrate inferior performance, since these reagents can interact strongly with the hydration layer of lipases due to their hydrophilic character. This leads to distortions in the hydration layer that may generate kinetic constraints in the reaction (reduction of catalytic activity) or inactivation of the biocatalyst (Banik et al., 2017; Bezbradica et al., 2007; Hari Krishna et al., 2001; Idris and Bukhari, 2012; Laane et al., 1987; Marty et al., 1997). Short-chain acids and alcohols are widely discussed in the literature as potent inhibitors of lipase activity (Claon and Akoh, 1993; Diaz et al., 2010; Couto et al., 2011; Gawas et al., 2016; Güvenç et al., 2002; Hari Krishna et al., 2001; Ikeda and Kurokawa, 2001; Kuperkar et al., 2014; Lotti et al., 2015; Sandoval et al., 2002; Sousa et al., 2022). The deleterious effects of the short-chain reagents (in this case, acetic acid and ethanol) can be attested by comparing the conversion and process metrics results in syntheses of ethyl acetate (both short-chain reagents) and n-octyl acetate (one short and one mid-chain reagent) (Tables 2 and 3). In contrast to the case of mid/long-chain ester syntheses, the reaction medium in the synthesis of short-chain esters becomes only moderately hydrophobic throughout the reaction, making the process more dependent on equilibrium shift strategies to achieve high conversions.

TABLE 3. Process metrics calculated from the alkyl esters enzymatic syntheses catalyzed by Novozym 435 (100 U biocatalyst loading) in equimolar condition (1:1 acid:alcohol) and stoichiometric excess (1:2 acid:alcohol) at 50 °C, 200 rpm, 4 h. All the syntheses were carried out using molecular sieves (30% wt/wt) to absorb the formed water.

The adoption of stoichiometric excess to shift the equilibrium toward synthesis led to a great increase in the conversion of n-octyl acetate synthesis (from 51.0% to 89.1%). In this case, the adoption of such a surplus quantity of n-octanol is justified. But this is not the case for other reactions (such as ethyl caprate synthesis) in which the stoichiometric excess of alcohol promoted only a discrete increase in conversion associated with a decrease in product concentration and volumetric productivity, as demonstrated in Table 3. One may argue that 1:2 represents 100% excess in relation to the limiting reagent, but literature often describes acid:alcohol molar ratios equal or greater than 1:3 (Matte et al., 2016; Remonatto et al., 2022b; Dai et al., 2014; Kuperkar et al., 2014; Pereira et al., 2018; Baek et al., 2020). Such stoichiometric excesses of alcohol generate effects like a solvent in synthesis, reducing the concentration of the acid and the formed product. This effect can be clearly seen when other process metrics are calculated, as shown in Table 3. Even the drastic increase in conversion in n-octyl acetate synthesis (48.1% increase), which is important to reduce the residual acidity in the media post-reaction, resulted in similar product concentration (only 0.8% increase) due to the stoichiometric excess of n-octanol (Table 3).

The mid/long-chain ester syntheses also demonstrate more favorable metrics when compared to the short-chain ones. Here, besides the thermodynamic and kinetics effects already discussed, the molar mass difference between short and mid/long-chain reagents and products also helps to explain the results. In parentheses, we included the process metrics on a molar basis to normalize such differences. We can check that the equimolar condition should be chosen in most cases, considering the metrics as a whole, even considering a reduction in reaction conversion. The ideal situation for the process is the achievement of a minimum quantity of unreacted reagents after reaction, which is obtained when we have quantitative conversion in equimolar (or near equimolar) condition, in a balance between short reaction time (space-time yield) and low (bio)catalyst loading (catalytic productivity). Quantitative conversion of reagents in equimolar condition will minimize the downstream operations required to purify the ester, which means saving in capital and operational expenditures for purification operations; high productivity will reduce the volume of the reactor for a given production scale, reducing the capital expenditure for its construction and the operational cost for its maintenance.

Although we obtained high product concentration in most cases, the volumetric productivities were not so high due to the reaction time (4 h). However, most results achieved 100 g L^-1 h^-1 or more, which is the minimum target suggested by Meissner and Woodley (2022) for low-priced products (<5.0 USD kg^-1). For optimization purposes we should balance conversion and product concentration, aiming for a reduction in reaction time. In a hypothetical case, if condition 1:1 of n-octyl caprate synthesis results in 80% (instead of 90.6% obtained) in 2 h reaction (instead of 4 h), the volumetric productivity of the process will increase from 0.65 mol L^-1 h^-1–1.18 mol L^-1 h^-1, an 81.5% increase. In summary, in assertive process development, these process metrics should be evaluated together (Meissner and Woodley, 2022).

To evaluate the catalytic productivity, we can use the propositions of Tufvesson and co-authors (2011) that suggested a minimum target of 400 g product g^-1 biocatalyst for chemical specialties production using immobilized biocatalysts. The results shown in Table 3 are the catalytic productivities obtained per reaction cycle, disconsidering the reuse of the same loading in successive cycles of reaction, one of the key features of immobilized biocatalysts. To achieve this minimum target in the synthesis’s studies here, we should achieve 10 or less cycles using the same loading for most cases. This target seems not far from the reported in the literature for reactions carried out in shaken flasks or CSTR-like reactors (Sousa et al., 2025) and can even be improved using proper reactors with less intensive mechanical stress on biocatalyst particles. However, if considered ester selling prices as bulk/commodities chemicals, the minimum target should be 2000 g g^-1 (Tufvesson et al., 2011), in which the reuse capacity of immobilized lipases should be high.

The lack of a process metrics-oriented approach in most literature studies can be attested in Table 4. Although most studies in the literature using Novozym 435 reported conversions superior to 90%, low productivity metrics were frequently observed, due to: i) long times without a proper evaluation of the kinetic profile of the reaction; ii) high biocatalyst loadings without a proper trade-off evaluation between biocatalyst loading, reaction time, and conversion; iii) high acid:alcohol molar ratios (Hauer, 2020). For instance, the work of Dai and co-authors (2014) evaluated reaction times between 12 and 36 h in SFS for butyl propionate synthesis and acid:alcohol molar ratios varying from 1:1 to 1:9, resulting in low STY. When present, solvents cause a reduction in product concentration and STY, as demonstrated in Table 4 (Baek et al., 2020; Rodriguez-Nogales et al., 2005; Jeromin and Zoor, 2008; Kuo et al., 2012). On the other hand, the study carried out by Bhavsar and Yadav (2018) evaluated the ethyl valerate synthesis in SFS activated by microwaves and, in this case, even a relatively low conversion (69.2%) in short-time (40 min) and equimolar condition resulted in high volumetric productivity (6.22 mol L^-1 h^-1 or 809.7 g L^-1 h^-1) (Bhavsar and Yadav, 2018). This case illustrates the need for a proper evaluation of reaction time and product concentration assessment over time in complement to the conversion results.

TABLE 4. Process metrics calculated from the alkyl esters enzymatic syntheses catalyzed by Novozym 435 described in the literature.

Synthesis in presence of solvents.

As the biocatalyst cost is elevated, the techno-economic feasibility of a biocatalytic process relies on the catalytic productivity. Catalytic productivity is dependent on biocatalyst loadings, and some works found optimal biocatalyst loadings as high as 5.0% wt/wt or more (Rodrigues-Nogalez et al., 2005; Kuperkar et al., 2014; Sá et al., 2018). Thus, even the achievement of maximum conversion using high loadings will result in relatively low catalytic productivity per cycle. In a previous work, we observed a relationship between high biocatalyst loadings and increased reuse capacity of immobilized lipases, but without proportionality – 5 times high Novozym 435 loading resulted in 3 times more reuses of the same loading (Sousa et al., 2022). An increase in catalytic productivity would be achieved if the opposite occurs – 5 times higher loading results in more than 5 times reuses, but this was not the case. Thus, the trade-off between biocatalyst loading and reuse capacity is not favorable for the process catalytic productivity, and consequently, to the process techno-economic feasibility.

Another important aspect related to process feasibility is the AVI, which should be as high as possible to increase the attractiveness of the industrial process. The relation between raw materials costs and product selling prices should be positive considering the process yield. AVI brings a fundamental task: the product selling price should be superior to the sum of raw material costs. This does not automatically indicates economic feasibility, but when AVI is equal to or inferior to one indicates that the proposed route is unfeasible. When the margin is low (low differences between costs and selling price), it is mandatory to reach the maximum yield. We surveyed some industrial and lab-scale prices of reagents and products to obtain an overview of the theoretical AVI, considering that the maximum conversion is achieved in the respective esterification reaction, for different esters production, as shown in Table 5. The lab-scale prices translated into an industrial-scale price context, divided by a factor of 20 as proposed by Tufvesson and co-authors (2011).

TABLE 5. Survey of market prices of carboxylic acids, alcohols, esters, and calculated added-value index (maximum theoretical value).

Echemi.

ChemCentral Univar.

Merck/Sigma-Aldrich.

We can check a great variation in prices between the different sources, particularly in the case of long-chain acids, such as oleic and stearic acids and the respective esters studied. This variation can be associated by the different purity grades or market availability. Some cases, on the other hand, show relative convergence in the theoretical AVI, such as ethyl acetate, butyl acetate, or n-octyl acetate. Among them, n-octyl acetate shows high AVI (between 2.0 and 3.0) considering two different sources of information.

n-Octyl acetate synthesis was chosen as a case study to further optimization, process design, and preliminary techno-economic assessment. We studied this synthesis in a bubble-column reactor prototype using a molar ratio acid:alcohol 1:2 and the biocatalyst loading 1.3%, at 50 °C. Bubble-column reactors (BCR) are described in the literature as an effective kind of reactor for enzymatic esterification using immobilized lipases (Liu et al., 2015; Baum et al., 2016), contributing with biocatalyst’s operational stability and kinetics rates improvement (Youssef et al., 2013). We achieved quantitative conversion (99.4) in 4 h using the BCR prototype. This result can be attributed to the efficient water removal by ascending bubbles and its absorption on the molecular sieves bed in the top of the reactor, and effective contact between the reagents and the biocatalyst in the media, favoring both thermodynamic and kinetics aspects of the reaction. In this condition, the same biocatalyst loading can be reused 9 times without any washing or post-reaction treatment. The same reuse capacity study was carried out in a stoppered flask, resulting in only three reuses without significant loss of enzymatic activity (inferior to 10% variation from the initial cycle conversion). Thus, in the BCR prototype, we observed an increase in process metrics compared to the stoppered flask–from 103.4 to 115.4 g L^-1 h^-1 STY and from 134.7 to 404.1 g g^-1 total catalytic productivity. The catalytic productivity achieved the minimum proposed by Tufvesson and co-authors (2011) for chemical specialties production using immobilized lipases. However, our survey (Table 5) indicates that n-octyl acetate selling price can be categorized as bulk/commodity (∼1 USD kg^-1), in which the minimum proposed is 2000 g g^-1.

The lab-scale data were used as input to a process design in SuperPro Designer, as demonstrated in Figure 1. The ChemSep simulator indicates that six distillation stages were required to separate all the unreacted n-octanol (that is recycled back to the reactor) from the n-octyl acetate, which is obtained pure (99%) at the end of the process. Table 6 indicates the mass balances and process metrics from a simulated production scale of ∼7000 metric ton year^-1 of n-octyl acetate.

TABLE 6. Mass balance, utilities, and process metrics calculation for octyl acetate continuous production (∼7000 metric ton yr^-1) using Novozym 435 as biocatalyst.

The data shown in Tables 5, 6 were used to calculate the economic parameters shown in Table 7, in which it is possible to observe that OPEX (USD 25.1 mi) is higher than the revenues on this scale (USD 16.6 mi), resulting in a negative Net Present Value (NPV): −64.2 million. This indicates the unfeasibility of the process in the conditions studied, in which the production cost per kg (4.09 USD kg^-1) is higher than the product selling price. The biocatalyst cost represents 74.3% of raw materials costs and 62.5% of total OPEX. In other words, the catalytic productivity of Novozym 435 for n-octyl acetate was not high enough to achieve feasibility considering the Novozym 435 cost (924 USD kg^-1) and the modest selling price of n-octyl acetate (2.41 USD kg^-1). Moreover, a reduction in both conversion of the reaction and catalytic productivity may be observed when scaling up, due to heat/mass transfer limitations in a large-scale scenario. Conversion reduction will increase the non-converted mass of acetic acid and n-octanol in the post-reaction stream, requiring additional distillation stages for recovery and purification of the ester, which means an increase in both CAPEX and OPEX for the same production scale. Catalytic productivity reduction, on the other hand, will require more biocatalyst, which means an increase in operational costs for the same production scale.

TABLE 7. Economic indicators for octyl acetate continuous production (∼7000 metric ton yr^-1) using Novozym 435 as biocatalyst.

In previous work, we studied the process design and techno-economic assessment of octyl caprylate (C16, long-chain ester) production using Lipozyme RM IM (lipase from Rhizomucor miehei immobilized) (Sousa and Ferreira-Leitão, 2026), in which we achieved 880.0 g product g^-1 biocatalyst due to the high reuse capacity of Lipozyme RM IM (20 sequential cycles) in the conditions tested. In this case, even the scenario with a lower selling price of ester (1.9 USD kg^-1), lower AVI (1.3), and higher biocatalyst cost (1000 USD kg^-1) compared to the case studied here, we obtained positive economic indicators: NPV 6.5 million USD and payback time 3.5 years, in a production scale of 10,000 ton yr^-1. This reinforces how crucial catalytic productivity is for the feasibility of the biocatalytic process.

The high-cost contribution of the biocatalyst in the studied process is not affected by the economies of scale effect, because additional biocatalyst mass is required when the production scale increases. Figure 2 demonstrated the sensitivity analysis of the NPV when varying the biocatalyst cost and the catalytic productivity (−100% to +100% variation).

In an optimistic scenario, in which conversion and catalytic productivity are equal in large-scale than the obtained in lab-scale, positive NPV is achieved when catalytic productivity is higher than 900 g g^-1, which means that the same loading must be reused at least 20 times (+113% variation, 2.3 times higher). Similarly, positive NPV is achieved when biocatalyst cost is reduced to less than 460 USD kg^-1 (−50% variation, 2 times lower). Thus, process optimization should focus on the reduction of biocatalyst cost contribution to the OPEX, which can be achieved by two different ways: i) reuse capacity increase or ii) adoption of a lower price biocatalyst with equivalent performance.

When considering only the acetic acid and n-octanol costs, we can observe that the stoichiometric excess of n-octanol led to high expenditure, which is offset by the recovery of n-octanol and is recycled back to the reactor. However, this recovery is possible due to the sequence of distillation columns, which contributes to CAPEX. CAPEX saving may occur using molar ratios lower than 1:2, but a trade-off between the residual acidity in post-reaction medium (that is high in equimolar condition when conversion decreases) and the surplus n-octanol to distillate (that is high in stoichiometric excess when conversion increases) should be carefully evaluated.

4 Conclusion

The study of reaction parameters guided by process metrics – product concentration, space-time yield, and catalytic productivity – can make the lab-scale study more robust, demonstrating the strengths and weaknesses of the process on a quantitative basis. We observed that mid- and long-chain ester syntheses exhibit favorable process metrics compared to short-chain ester syntheses due to the high initial hydrophobicity of the media. This reaction media feature lead to the achievement of high conversion in equimolar condition, resulting in high product concentrations. Considering the magnitude of differences between the immobilized lipase (Novozym 435) and esters prices, catalytic productivity becomes crucial in the development of a potentially feasible biocatalytic process. We demonstrate that literature dedicates less attention to the impact of reaction conditions (molar ratio, biocatalyst loading/reaction time) on the process metrics. Furthermore, we translated the reaction conditions of n-octyl acetate synthesis in lab-scale to simulated large-scale scenarios, in which this impact was even more evident, demonstrating the challenges to apply biocatalysis for producing commodities/specialty chemicals. The process metrics evaluation helps to understand the bottlenecks of the process in numbers, leading to a concise definition of paths to overcome them.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Pablo Domínguez de María, Sustainable Momentum, SL, Spain

Reviewed by: David Gonzalez Miranda, Institut National des Sciences Appliquées de Toulouse (INSA), France

Aditya Sarnaik, Arizona State University - Polytechnic Campus, United States